Skip to content

v0.17.0

Choose a tag to compare

@github-actions github-actions released this 30 Oct 06:35
· 3 commits to main since this release
d7ebd06

A MINOR: memory_read's tools/list entry gains a _meta field. The other
changes are tool description text and documentation.

Added

  • memory_read is kept loaded up front in Claude Code. Its tools/list
    entry carries _meta: {"anthropic/alwaysLoad": true}. Claude Code puts MCP
    tools behind tool search by default, so at the start of a task the read could
    be out of context entirely; this flag (Claude Code 2.1.285,
    https://code.claude.com/docs/en/mcp.md) opts one tool out. Only memory_read
    carries it; every other tool stays behind search. The flag takes effect once
    the server has connected: per the same page, "Startup doesn't wait for your
    server because of this value", and Claude Code connects servers in the
    background by default, so a first prompt can still start before the tool
    list arrives (MCP_CONNECTION_NONBLOCKING=0 makes startup wait). Clients
    that do not know the key ignore _meta. No tool text changes: the flag is
    metadata, and the descriptions stay as reviewed. Pinned in
    test/always-load.test.ts.

Changed

  • Make memory_read, memory_write and memory_create_room descriptions agent-agnostic for Anthropic Directory review. memory_read now says that a correction is a fresh memory_write, stored alongside the earlier memory (which is not deleted or hidden) unless it is too close to an existing memory, and that no MCP deletion tool is exposed: a user can permanently delete an individual personal memory through the REST API or with help from support. memory_create_room explains shared-room history access and private-memory isolation. Tool behavior, schemas, permissions and defaults are unchanged.

Documentation

  • The README names the official sources. This repository, the npm
    package and the .mcpb attached to GitHub Releases; a copy of the README
    elsewhere with its own "Download" button is not ours (#233).

  • Every client starts with the keyless path; the API key has its own
    section.
    The README went from a short no-key block straight to "Get a free
    API key", and showed Claude Desktop, Devin Desktop (formerly Windsurf), Zed,
    JetBrains and Cline only with MNEMOVERSE_API_KEY; npm, GitHub and the
    directories that copy the first mcpServers block passed that on. The Quick
    Start now connects every client to the hosted server
    https://mcp.mnemoverse.com/mcp with a browser sign-in: natively where the
    client adds a remote URL (Claude Code, Cursor, VS Code, a Claude Desktop
    custom connector, Cline), and through the mcp-remote bridge
    (npx -y mcp-remote@latest https://mcp.mnemoverse.com/mcp) where it only
    starts local commands (Claude Desktop's config file, Devin Desktop, Zed,
    JetBrains). The README's first mcpServers block is that bridge entry, with
    no key. "Local package with an API key" follows as its own section, with the
    same key facts as before and one more: a client that can send a header can
    use the hosted URL with Authorization: Bearer mk_live_.... Seven keyless
    partials, docs/snippets/<client>-hosted.md, sit beside the API-key ones;
    the five that mnemoverse.com/docs mirrors are unchanged. llms-install.md
    and llms.txt give Cline's remote entry and the bridge before the local
    steps.

  • npm: https://www.npmjs.com/package/@mnemoverse/mcp-memory-server

  • MCP Registry: https://registry.modelcontextprotocol.io/v0.1/servers?search=mnemoverse

What's Changed

  • feat(tools): keep memory_read loaded up front in Claude Code (anthropic/alwaysLoad) by @OlyaTi in #234
  • docs(readme): name the only official sources by @izgorodin in #233
  • docs(install): every client starts keyless; the API key gets its own section by @izgorodin in #232
  • Clarify memory controls and neutralize Directory tool descriptions by @izgorodin in #215
  • chore(release): 0.17.0 by @izgorodin in #236

Full Changelog: v0.16.0...v0.17.0